基于BP神经网络的大视场成像畸变校正方法

史泽林; 康 娇; 孙 锐

doi:null

您当前的位置：

首页 >

文章列表页 >

基于BP神经网络的大视场成像畸变校正方法

更新时间：2020-08-12

- 基于BP神经网络的大视场成像畸变校正方法
- BP NN-based method for lens distortion correction of large-field imaging
- 光学精密工程 2005年13卷第3期页码：348-353
- 作者机构：
  
  1. 中国科学院沈阳自动化研究所,辽宁沈阳,110016
  2. 中国科学院研究生院北京,100039
- 作者简介：
- 基金信息：
- DOI：
  中图分类号： TP183
- 收稿日期：2005-03-22，
  
  修回日期：2005-04-18，
  
  网络出版日期：2005-06-30，
  
  纸质出版日期：2005-06-30
- 稿件说明：
移动端阅览
史泽林, 康娇, 孙锐. 基于BP神经网络的大视场成像畸变校正方法[J]. 光学精密工程, 2005,13(3):348-353.

SHI Ze-lin, KANG Jiao, SUN Rui. BP NN-based method for lens distortion correction of large-field imaging[J]. Optics and precision engineering, 2005, 13(3): 348-353.
史泽林, 康娇, 孙锐. 基于BP神经网络的大视场成像畸变校正方法[J]. 光学精密工程, 2005,13(3):348-353. DOI：

SHI Ze-lin, KANG Jiao, SUN Rui. BP NN-based method for lens distortion correction of large-field imaging[J]. Optics and precision engineering, 2005, 13(3): 348-353. DOI：

摘要

提出了基于BP神经网络的畸变校正方法

实现了在不确知畸变数学模型情况下的高精度校正。应用平行光管法测得900组光点坐标数据作为BP神经网络的训练样本

通过训练得到一个隐含层有6个节点的3层BP神经网络模型

然后使用77组数据检测了该神经网络的校正效果

并与使用相同检测数据的双线性插值方法进行了比较。实验表明

神经网络较插值方法精度高

误差数据分布的规律性好

而且与插值方法相比

它避免了误判区域的问题

便于校正。

Abstract

A method based on BP neural network(NN) is proposed in this paper

which realizes high precision of correction without knowing the mathematic model. 900 groups of light-point data are obtained as training samples by parallel light pipe method. A three-layer NN model with 6 nodes in the hidden layer is achieved and its correction effect is examined by 77 groups of data. The dual linear interpolation method is applied to the same 77 groups of data in comparison. The conclusion is that comparing to interpolation method

the BP NN method has higher precision and the BP NN method can avoid the problem of error estimation region in the interpolation method.

关键词

Keywords

references

. TIAN H,SRIKANTHAN T,VIJAYAN K,et al. Study on the effect of object to camera distance on polynomial expansion coefficients in barrel distortion correction . Fifth IEEE Southwest Symposium on Image Analysis and Interpretation,Santa Fe, New Mexico,2002.

. CUCCHIARA R,GRANA C,PRATI A,et al.A Hough transform-based method for radial lens distortion correction . Proceedings of the 12th International Conference on Image Analysis and Processing,Mantova, Italy,2003.

. ZHANG G J,HE J J,YANG X M.Calibrating camera radial distortion with cross-ratio invariability[J]. Optics and Laser Technology,2003,35:457-461.

. 张华达. 大视场短焦距CCD成像系统畸变校正研究[J]. 红外,2004,1(1):1-6. ZHANG H D. Research on distortion correction of large-field and short-focus CCD imaging system[J].HongWai,2004,1(1):1-6.(in Chinese)

. 行麦铃,刘贱平. 大视场短焦距镜头CCD摄像系统的畸变校正[J].光学技术,2003,29(3):377-379. XING M L,LIU J P.Distortion measurement of CCD camera with a large field, short focal length lens[J] .Optical Technique,2003,5(1):377-379.(in Chinese)

. 刘金国. 大视场光电测量系统的精密几何标定和畸变校正的研究[J]. 光学精密工程,1994,2(4):109-120. LIU J G.Study on high accurate geometric calibration of precision photoelectronic measuring system with large field of view and distortion correcting[J]. Optics and Precision Engineering,1994,2(4):109-120.(in Chinese)

. 李英才,徐军,田少文. 大视场CCD测角的实时误差修正 . 光子学报,1993,22(4):446-450. LI Y C,XU J,TIAN SH W.Real-time error amending of large-field CCD measuring angle[J]. Acta Photonica Sinica,1993,4(22):446-450.(in Chinese)

. BURKHARDT S,SCHWEIKARD A,BURGKART R. Numerical determination of the susceptibility caused geometric distortions in magnetic resonance imaging[J]. Medical Image Analysis,2003,7:221-236.

. 李德强.基于小波网络与模糊神经网络的非线性系统辨识方法的研究 . 中国科学院沈阳自动化研究所,2002. LI D Q.Research on identification method for nonlinear system based on wavelet networks and fuzzy neural networks . PhD. Dissertation, Shenyang Institute of Automation, Chinese Academy of Sciences,2002.(in Chinese)

浏览量

304

下载量

CSCD

文章被引用时，请邮件提醒。

提交

工具集

关联资源

高频磁致伸缩换能器输出特性测试与优化控制

精密减速器检测仪测角误差补偿

高速扫描激光共聚焦显微内窥镜图像校正

改进的空间大口径追踪相机畸变校正

基于Pearson关联度BP神经网络的时间序列预测